[Review Shape] CLDepthwiseConvolutionLayer mismatches

- Fixed a bug that corrected the number of dimensions of a TensorShape for added trailing 1s
- Avoided adding offset_first_element for the Depthwise 3x3 NCHW OpenCL kernels, since it wouldn't align with the window which is based on the output
- Adjusted padding requirements along the x for Depthwise 3x3 NCHW. The kernel should always add 2 * dilation_(x/y) to the num_elems_read_x/y
- Adjusted the kernel's border_size given to the border handler at function level
- Added the dataset that previously made the tests fail

Resolves: COMPMID-4041

Change-Id: Ifab7d38b263f12173fcc96a5f0bd3375756c3c53
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4673
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: SiCong Li <sicong.li@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>

1 files changed, 5 insertions, 4 deletions

diff --git a/tests/datasets/DepthwiseConvolutionLayerDataset.h b/tests/datasets/DepthwiseConvolutionLayerDataset.h
index ed596d6d45..86804fb4c6 100644
--- a/tests/datasets/DepthwiseConvolutionLayerDataset.h
+++ b/tests/datasets/DepthwiseConvolutionLayerDataset.h
@@ -161,9 +161,9 @@ class SmallDepthwiseConvolutionLayerDataset3x3 final : public DepthwiseConvoluti
 public:
     SmallDepthwiseConvolutionLayerDataset3x3()
     {
-        add_config(TensorShape(3U, 3U, 2U), Size2D(3U, 3U), PadStrideInfo(1, 1, 0, 0));
+        add_config(TensorShape(1U, 1U, 2U), Size2D(3U, 3U), PadStrideInfo(1, 1, 2, 2));
         add_config(TensorShape(7U, 8U, 3U, 2U), Size2D(3U, 3U), PadStrideInfo(1, 1, 0, 0));
-        add_config(TensorShape(21U, 31U, 9U, 4U), Size2D(3U, 3U), PadStrideInfo(1, 1, 1, 0));
+        add_config(TensorShape(32U, 31U, 9U, 4U), Size2D(3U, 3U), PadStrideInfo(1, 1, 0, 0));
         // Asymmetric padding
         add_config(TensorShape(33U, 27U, 11U), Size2D(3U, 3U), PadStrideInfo(2, 2, 0, 1, 0, 1, DimensionRoundingType::FLOOR));
     }
@@ -186,7 +186,6 @@ class LargeDepthwiseConvolutionLayerDataset3x3 final : public DepthwiseConvoluti
 public:
     LargeDepthwiseConvolutionLayerDataset3x3()
     {
-        add_config(TensorShape(33U, 27U, 11U, 3U), Size2D(3U, 3U), PadStrideInfo(1, 1, 0, 1));
         add_config(TensorShape(33U, 27U, 11U, 3U), Size2D(3U, 3U), PadStrideInfo(1, 1, 1, 1));
         add_config(TensorShape(21U, 31U, 9U, 4U), Size2D(3U, 3U), PadStrideInfo(1, 2, 1, 0));
         add_config(TensorShape(33U, 27U, 11U, 3U), Size2D(3U, 3U), PadStrideInfo(1, 2, 0, 1));
@@ -202,6 +201,8 @@ public:
         add_config(TensorShape(233U, 277U, 55U), Size2D(3U, 3U), PadStrideInfo(1, 2, 0, 0));
         add_config(TensorShape(333U, 277U, 77U, 5U), Size2D(3U, 3U), PadStrideInfo(2, 3, 0, 1));
         add_config(TensorShape(177U, 311U, 22U), Size2D(3U, 3U), PadStrideInfo(2, 1, 1, 1));
+        // Width and height are a multipile of the processing tile size
+        add_config(TensorShape(32U, 21U, 11U, 3U), Size2D(3U, 3U), PadStrideInfo(1, 1, 0, 1));
     }
 };
 
@@ -269,4 +270,4 @@ public:
 } // namespace datasets
 } // namespace test
 } // namespace arm_compute
-#endif /* ARM_COMPUTE_TEST_DEPTHWISE_CONVOLUTION_DATASET */
+#endif /* ARM_COMPUTE_TEST_DEPTHWISE_CONVOLUTION_DATASET */
\ No newline at end of file